Google Maps and Live Big Data

I like to bore my wife with non-essential trivia and she indulges me by pretending to be twice as interested as she actually is.

Last Sunday we were on our way back from a long weekend holiday with extended family. We hit way less traffic than we did on the way over the Wednesday before. Only the expected beach traffic consisting out of Belgians and Germans on their way to enjoy a warm and sunny Sunday on the Dutch beaches. But while I was patiently merging with all the other cars to take the exit, I kept a watchful eye on Google Maps to see if my ETA would change. And that reminded me of a little trivia of the navigation features of Google Maps. The way Google estimates your Estimated Time of Arrival.

Estimating the time to travel from A to B is—partially—pretty straightforward. Google does the basic math equation where it calculates how long it takes to drive X kilometres when the car is going X kmph. They combine that with information about the speed limits and historic data of previous users driving that route, or pieces of it. But how can it tell me what my delay will be while I'm inside the traffic jam? How can it recommend a new faster route while I'm only moments away from hitting a congested part around Antwerp?

It collects that live from other users. Uhu, that's right. If somebody else somewhere in front of you is also using Google Maps to find their way, that data will be used to fine-tune the algorithm that makes the estimation. Your information is likewise used for people that are close behind you.

I'm not sure though if that still happens when you disable location history. That location history thing made headlines not long ago, at the height of the privacy craze (not to diminish the legitimacy of those worries) with Facebook scandals popping up left and right. It became known to the wider public, me included, that Google still tracks you even if you have location sharing turned off. Although I'm not sure, but to me that sounds like the service Google uses to give those accurate estimations.

The live tracking isn't always perfect. When a traffic jam is just starting, it's hard to know how long the delay will be. Cars already in that jam have not been from back to front yet. On two occasions, like last Wednesday, I left with an estimated travel time that for the next 20 to 30 minutes did not change. Conjunctions were only just forming and my estimated travel time was increasing at the same rate as I was traveling. You can also notice the imperfection on the parts of the route that Google colors orange. Those orange bits, which mean slow traffic, also show on parts of the route that has traffic lights. Or in my case last weekend, a bit of road where I was allowed 50kmph but had these speed bumps that were best taken at less than 20kmph.

A necessary evil

Google needs that data to make accurate predictions. The more data they collect, the better the service is they can provide. And the better the service is that they provide, the more the user is expecting from the service or any competing services. But to provide the level of service their user is expecting or to provide a better service than their competition, Google needs to be this invasive. Google Assistent is (or perhaps was because I don't really know the current state of things) considered better than Siri. That is because it's more intrusive than their Apple counterpart.

Think of it like a butler or any of the other house staff you can see on Downton Abby. They know an awful lot about the people they serve. But that's the cost of having such help. During the era Downton Abby depicts, the ability to keep their mouth shut was highly valued. The house staff's livelihoods depended on that status if they ever needed to work somewhere else. But to best help their employers the house staff needed to be there for the most intimate moments. They couldn't help but to get to learn their employer very well. Their employer expected a certain level of service. The best butler is the butler that knows you through and through and is able to anticipate on your needs.

You can programme a Nest thermostat to heat up the house at certain times, like when you regularly get home.
But what if you have an irregular schedule? Sure, you can turn it on manually from a distance, but wouldn't it be nice if it could turn on when it noticed you were on your way home. That would require it to know your location and that of any other household member at any time. And it should have enough knowledge about your habits to know if your current movement means you're coming home. You would have to give up quite some personal details.

Perhaps one day, AI advancements will be far enough to have us enjoy a digital personal assistant or thoughtful products and services, without relinquishing personal data or surrendering our privacy. But for now we'll have to make do with what we have and be mindful about what information we share and what not. Finding a balance between service and privacy.